Abstract
Most previous work investigating the relationship between obesity and food choice in adolescents has focused on single foods or food groups. In contrast examining diet in its entirety, by way of dietary patterns, might allow for more insight and understanding in the relationship between obesity and food choice. The main aim of this thesis was to develop and investigate dietary patterns in older adolescents.
A cross-sectional study of 693 adolescents aged 15-18 years from Otago secondary schools was conducted. The students completed an online survey, which collected information on demographics, food intake, and dieting status. Height, weight, bio-electrical impedance and waist circumference were obtained. A subset of 271 participants completed the 20-metre multistage fitness test to assess cardiorespiratory fitness; these data were converted to ⩒O2max. Principal components analysis (PCA) was used to determine dietary patterns. Regression analysis was used to examine associations, with robust standard errors and accounting for schools as clustering units.
PCA produced three dietary patterns: ‘Treat Foods’, ‘Fruits and Vegetables’, and ‘Basic Foods’. Every school day breakfast was eaten the ‘Treat Foods’ score decreased by 0.08 (95% CI: -0.13, -0.03) and during the weekend by 0.17 (95% CI: -0.28, -0.05). Weekday and weekend breakfast and dinner frequency were positively associated with the ‘Fruits and Vegetables’ pattern. Weekend breakfast and dinner frequency were positively associated with the ‘Basic Foods’ pattern. These results suggest increased frequency of eating meals provides opportunity for a variety of foods to be consumed, and may reduce the consumption frequency of unhealthy snack foods increasing the likelihood of achieving a “healthy” dietary intake.
Significant associations were found between the PCA patterns overall and percent body fat, fat-to-lean ratio and the fat mass index (FMI). Every unit increase in the ‘Fruits and Vegetables’ score was associated with a 2.2% decrease in the geometric mean of percent body fat. For every unit increase in the ‘Basic Foods’ score an expected decrease of 2.5%, 3.1%, and 3.6% was seen in the geometric mean of percent body fat, fat-to-lean ratio and the FMI, respectively. A significant current dieting status by ‘Treat Foods’ interaction was found for body mass index (BMI), waist circumference and waist-to-height ratio. As significant associations were found between dietary patterns and indices of both central and total adiposity, but not BMI, using only BMI may not be useful in this age group.
Fitter individuals were found to have higher ‘Fruits and Vegetables’ pattern scores. Girls with both low and high ⩒O2max had higher ‘Treat Food’ pattern scores than those girls with moderate cardiorespiratory fitness, while a negative association was seen in the boys. These results suggest targeting both cardiorespiratory fitness and food choice as public health messages may be important.
This research has demonstrated that the PCA derived dietary patterns are likely valid predictors of health status of Otago adolescents.